sathiiiii/polyalign-gemma2-2b-en-dist-sft

TEXT GENERATIONConcurrency Cost:1Model Size:2.6BQuant:BF16Ctx Length:8kPublished:Apr 21, 2026License:otherArchitecture:Transformer Cold

The sathiiiii/polyalign-gemma2-2b-en-dist-sft model is a 2.6 billion parameter language model, fine-tuned from Google's Gemma-2-2B architecture. This model has been specifically fine-tuned on the polyalign_dist_sft_train dataset, indicating an optimization for specific supervised fine-tuning tasks. It is designed for applications requiring a compact yet capable model within the Gemma 2 family.

Loading preview...

Model Overview

The sathiiiii/polyalign-gemma2-2b-en-dist-sft is a 2.6 billion parameter language model derived from the google/gemma-2-2b base architecture. This model has undergone supervised fine-tuning (SFT) using the polyalign_dist_sft_train dataset, suggesting a specialization for tasks aligned with this training data.

Key Training Details

  • Base Model: google/gemma-2-2b
  • Fine-tuning Dataset: polyalign_dist_sft_train
  • Parameters: 2.6 billion
  • Context Length: 8192 tokens
  • Learning Rate: 1e-05
  • Optimizer: ADAMW_TORCH_FUSED
  • Epochs: 1.0
  • Mixed Precision: Native AMP

Performance Metrics (on evaluation set)

During training, the model achieved a validation loss of 1.2383. The training process involved 9000 steps across 1.0 epoch, with a total batch size of 64 across 8 devices.

Intended Use Cases

While specific intended uses and limitations are not detailed in the provided README, its fine-tuned nature suggests suitability for tasks similar to those present in the polyalign_dist_sft_train dataset. Developers should evaluate its performance for their specific applications, particularly where a compact Gemma 2-based model with specialized SFT is beneficial.